Univerzitní knihovna OU - Úplné zobrazení záznamu

Intro -- Abstract -- Acknowledgments -- List of Publications -- Contents -- I Comprehensive Summary -- 1 Introduction -- 1.1 Research Scope and Delimitations -- 1.2 Thesis Overview -- 2 Background -- 2.1 Machine Learning Basics -- 2.2 Digital Pathology -- 2.3 Artificial Intelligence in Pathology -- 3 Theoretical Framework -- 3.1 Automation and Human Control -- 3.2 Challenges Designing with ML -- 3.3 AI-assisted Decision-making -- 4 Research Approach -- 4.1 Constructive Design Research -- 4.2 Motivations -- 4.3 From Design Experiments to Generative Knowledge -- 4.4 Designing for Efficiency -- 5 Paper summary -- 6 Design Experiments -- 6.1 Overview -- 6.2 Exploring Proactive Training Data Collection (DROID) -- 6.3 AI-assisted Annotation (TW) -- 6.4 AI-assisted Visual Search (LGL) -- 6.5 AI-assisted Quantification (PDL1) -- 7 A Framework for Designing Human-Centred Machine Learning -- 7.1 The Importance of Thoughtful Action -- 7.2 Three Interconnected Activities -- 7.3 The Impact of Gathering Training Data -- 7.4 The Impact of Interaction Design -- 7.5 The Impact of Model Development -- 8 Conclusion and Discussion -- 8.1 Summary of Contributions -- 8.2 Why is Designing with ML Difficult? -- 8.3 Power to the People? Reflections on Interactive Machine Learning -- 8.4 The Role of Constructive Design Research -- 8.5 In Conclusion -- Bibliography -- II Appended papers -- 1 TissueWand, a Rapid Histopathology Annotation Tool. -- 2 Rapid Assisted Visual Search: Supporting Digital Pathologists with Imperfect AI. -- 3 From Machine Learning to Machine Teaching: The Importance of UX. -- 4 Machine Learning as a Design Material: a Curated Collection of Exemplars for Visual Interaction. -- 5 Verification Staircase: a Design Strategy for Actionable Explanations. -- 6 Designing for the Long-Tail of Machine Learning.

Anotace

Annotation